4:10 PM - 4:30 PM
[3F5-GS-10-03] Prediction of Liner Shipping Operations by Machine Learning Using AMeDAS Data
Keywords:operation forecast, AMeDAS, machine learning, SVM
In Japan, where many islands are remote, access to them is often by liner ship, and deciding whether to depart is essential for daily life and tourism. For example, on Tobishima Island in Sakata City, Yamagata Prefecture, the decision to depart is made on the morning of the day based on meteorological data such as wind speed and wave height in the surrounding area. Therefore, it is difficult to make a plan for the departure of a liner because it is yet to be known whether or not the liner will be able to depart until the day of departure. In this study, we developed a support vector machine (SVM) system for predicting liner vessel operations for the next day and beyond based on publicly available automated meteorological data acquisition system (AMeDAS) data from the Japan Meteorological Agency and ten years of actual liner ship operations. By combining AMeDAS data from multiple locations, the system achieved a prediction accuracy of 72% for one week. In addition, we have developed a system that provides a one-week forecast of liner ship operations via the Web.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.